EP3352415B1

EP3352415B1 - Smb service failure handling method, and storage device

Info

Publication number: EP3352415B1
Application number: EP16910788.5A
Authority: EP
Inventors: Zhengjiang Wang; Xianqiang Luo; Yi He; Keji HUANG; Jian Wu; Mingqian Zhang
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2016-11-25
Filing date: 2016-11-25
Publication date: 2021-05-19
Anticipated expiration: 2036-11-25
Also published as: CN108352995B; US10789138B2; CN108352995A; US20180225183A1; EP3352415A1; WO2018094686A1; EP3352415A4

Description

TECHNICAL FIELD

The present invention relates to the communications field, and in particular, to an SMB service fault processing method and a storage device.

BACKGROUND

A server message block (SMB) protocol is a protocol for sharing a file, a printer, or a serial port between computers. By using the SMB protocol, a client may read/write a file on a server in various network environments, and may further make a service request to the server. In a Transmission Control Protocol/Internet Protocol (TCP/IP) environment, after the client establishes a connection to the server, the client may send an SMB command to the server, and operate a file system, for example, access a shared directory, open a file, or read/write a file.
In a process of performing an input/output (I/O) port-based operation on a file, an application program needs to invoke a file operation function and transmit a file name, and further needs to select a path to the file to open the file, so as to read data from the file. The function is used to retrieve a sequence number, that is, a file handle, and the file handle is a unique identification basis for the opened file. The application program needs to invoke a read file function, and transmit the file handle and a specified byte count to an operating system, so as to read a piece of data from the file.
In the prior art, when a host accesses a file in a first storage device, a file system of the first storage device generates a file handle according to a file open instruction sent by the host, and opens the file by using the file handle. If the first storage device is faulty, when the first storage device hands over to a second storage device, a file system of the second storage device cannot respond to a file open instruction resent by the host because the second storage device does not have the file handle. Consequently, the host cannot access the file, and an SMB service is interrupted.
EP 3 051 420 A1 discloses an SMB service of fault processing method providing transparent failover capabilities by suggesting the use of persistent handles which are replicated between nodes of the file system.

SUMMARY

The present invention provides an SMB service fault processing method and a storage device, so that when a storage device that processes an SMB service is faulty, the SMB service can be taken over by a secondary storage device, so as to ensure that the SMB service is executed without interruption.
The invention is defined by the SMB service fault processing method of claim 1 and a storage system as defined in claim 8. Additional features of the invention are provided in the dependent claims. In the following, parts of the description and drawings referring to embodiments, which are not covered by the claims, are not presented as embodiments of the invention, but as examples useful for understanding the invention.
A first aspect of the present invention provides an SMB service fault processing method, where the method may be applied to a service processing system, and the service processing system includes a host and at least two storage devices. The method includes: receiving, by a first storage device, a first file open instruction sent by the host; generating a file handle according to the first file open instruction, sending the file handle to a second storage device, and sending a handle identifier of the file handle to the host; saving, by the second storage device, the file handle as a backup handle; when the first storage device is faulty, establishing, by the host, a communication connection to the second storage device; and determining, by the second storage device according to the handle identifier in a second file open instruction sent by the host, the file handle corresponding to the handle identifier, opening a file according to the file handle, and sending an operation result to the host, where the first storage device is a service device, and the second storage device is a backup device. In this implementation, after the first storage device backs up the file handle to the second storage device, when the first storage device is faulty, the second storage device may recover an SMB service according to the file handle, so as to ensure that the SMB service is executed without interruption.
In a possible implementation of the first aspect, the first storage device generates the file handle according to a persistent handle obtaining subcommand included in the first file open instruction; checks whether the file handle is a persistent handle; and if the file handle is a persistent handle, sends the file handle to the second storage device. In this implementation, the second storage device backs up only a persistent handle.
In another possible implementation of the first aspect, the second file open instruction includes a handle restoration subcommand, and the handle identifier is carried in the handle restoration subcommand.
In another possible implementation of the first aspect, when the first storage device, the second storage device, and a distributed lock server belong to a same storage array, before sending the file handle to the second storage device, the first storage device receives a lock instruction sent by the host, and sends a file lock request to the distributed lock server according to the lock instruction; the distributed lock server sets, according to the file lock request, a file lock for the file corresponding to the file handle; and then the first storage device sends the file lock to the second storage device; and the second storage device stores the file lock as a backup file lock, where the first storage device may further establish a correspondence between the file lock and the file handle. In this implementation, in addition to the file handle, the secondary storage device may further back up the file lock.
Further, in another possible implementation of the first aspect, after storing the file handle as the backup handle, the second storage device sends a silence request to the distributed lock server, where file information carried in the silence request is corresponding to the backup handle; and the distributed lock server silently sets, according to the file information, a file corresponding to the backup handle in the second storage device. In this implementation, after silently setting the file, the distributed lock server does not allocate a file lock for the file when receiving another lock request for the file.
Further, in another possible implementation of the first aspect, after determining the file handle corresponding to the handle identifier, the second storage device determines the backup file lock corresponding to the file handle, and sends information about the backup file lock to the distributed lock server; and the distributed lock server locks the file according to the information about the backup file lock.
In another possible implementation of the first aspect, when the first storage device and a first distributed lock server belong to a same storage array, and the first storage device and the second storage device belong to different storage arrays, before sending the file handle to the second storage device, the first storage device receives a lock instruction sent by the host, and sends a file lock request to the first distributed lock server according to the lock instruction; the distributed lock server sets, according to the file lock request, a file lock for the file corresponding to the file handle; the first storage device sends the file lock to the second storage device; and the second storage device stores the file lock as a backup file lock. In this implementation, in addition to the file handle, the secondary storage device may further back up the file lock.
Further, in another possible implementation of the first aspect, a second distributed lock server and the second storage device belong to a same storage array. After storing the file handle as the backup handle, the second storage device sends a silence request to the second distributed lock server, where file information carried in the silence request is corresponding to the backup handle; and the second distributed lock server silently sets, according to the file information, a file corresponding to the backup handle in the second storage device. In this implementation, when the first storage device and the second storage device belong to different storage arrays, after silently setting a backup file, the distributed lock server in a second storage array does not allocate a file lock for the backup file when receiving another lock request for the file.
Further, in another possible implementation of the first aspect, after obtaining the file handle corresponding to the handle identifier, the second storage device determines the backup file lock corresponding to the file handle, and sends information about the backup file lock to the second distributed lock server; and the second distributed lock server locks the file according to the information about the backup file lock. In this implementation, in addition to the file handle, the secondary storage device may further restore the file lock for the file, so as to ensure integrity of an SMB service.
Further, in another possible implementation of the first aspect, the second storage device starts timing when the first storage device is faulty; and when a timing duration reaches a preset duration, the second storage device deletes the file handle, or sets the file handle to be invalid. In this implementation, when the service device is provided with multiple secondary storage devices, the secondary storage device may set an aging duration. When the aging duration expires, the secondary storage device may clear the file handle, so as to save storage resources.
A second aspect provides an SMB service fault processing method, including: sending, by a host, a first file open instruction to a first storage device, where the first file open instruction includes a persistent handle request; when the first storage device is faulty and an IP address of the first storage device is activated on a second storage device, sending, by the host, a second file open instruction to the second storage device, where the second file open instruction includes a handle identifier, and the handle identifier is used to instruct the second storage device to obtain a file handle of a file; and receiving, by the host, an operation result returned by the second storage device. It can be learned that when the service device is faulty, the host may execute an SMB service on the backup device, and the backup device may read/write an SMB service file according to the file handle, so as to ensure that the SMB service is executed without interruption.
In a possible implementation of the second aspect, the first file open instruction includes a persistent handle obtaining subcommand, and the persistent handle obtaining subcommand is used to instruct the first storage device to generate a persistent handle.
In another possible implementation of the second aspect, the second file open instruction includes a handle restoration subcommand, and the handle restoration subcommand carries the handle identifier.
In another possible implementation of the second aspect, the host sends a lock instruction to the first storage device, where the lock instruction is used to set a file lock for the file corresponding to the file handle.
A third aspect provides a storage system, including a first storage device and a second storage device. The first storage device may implement a function of the first storage device in the SMB service fault processing method provided in the first aspect. The second storage device may implement a function of the second storage device in the SMB service fault processing method provided in the first aspect. A function of a storage device may be implemented by hardware, or may be implemented by executing corresponding software by hardware. A fourth aspect provides a host, which may implement a function of the host in the SMB service fault processing method provided in the second aspect. A function of a host may be implemented by hardware, or may be implemented by executing corresponding software by hardware.
A fifth aspect provides a service processing system, including the host provided in the fourth aspect and the storage system provided in the third aspect.
It can be learned from embodiments of the present invention that after generating the file handle according to the first file open instruction sent by the host, the first storage device sends the file handle to the second storage device; the second storage device saves the file handle as the backup handle; and when the first storage device is faulty, the second storage device establishes the communication connection to the host, receives the second file open instruction sent by the host, determines, according to the handle identifier in the second file open instruction, the file handle corresponding to the handle identifier, opens the file according to the file handle, and sends the operation result to the host. In the embodiments of the present invention, the file handle may be backed up in a different storage array or in a different storage device in a same storage array. When the service device (that is, the first storage device) is faulty, the backup device (that is, the second storage device) may take over the SMB service, so as to ensure that the SMB service is executed without interruption.

BRIEF DESCRIPTION OF DRAWINGS

FIG. 1 is a diagram of an architecture of a service processing system according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a storage device according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of an SMB service fault processing method according to an embodiment of the present invention;
FIG. 4 is another schematic diagram of an SMB service fault processing method according to an embodiment of the present invention;
FIG. 5 is another schematic diagram of an SMB service fault processing method according to an embodiment of the present invention;
FIG. 6 is another schematic diagram of an SMB service fault processing method according to an embodiment of the present invention;
FIG. 7 is a schematic structural diagram of a service processing system according to an embodiment of the present invention;
FIG. 8 is another schematic structural diagram of a service processing system according to an embodiment of the present invention;
FIG. 9 is another schematic structural diagram of a service processing system according to an embodiment of the present invention; and
FIG. 10 is a schematic structural diagram of a host according to an embodiment of the present invention.

DESCRIPTION OF EMBODIMENTS

The present invention provides an SMB service fault processing method, and the method is applied to a service processing system. Referring to FIG. 1, FIG. 1 is a schematic diagram of an architecture of a service processing system according to an embodiment of the present invention.
The service processing system includes a host 11 and multiple storage arrays 12.
The host 11 is a computing device that has a computing capability, and can complete various types of data processing work. The host may be a personal computer (PC), a mobile phone, a tablet, an in-vehicle computer, a wearable electronic device, a personal digital assistant (PDA), or the like.
The storage array 12 includes one or more storage devices 13. The storage device 13 has one or more backup devices. These backup devices may be located in the same storage array as the storage device 13, or may be located in a different storage array from the storage device 13. For example, in a first storage array, a storage device B is a backup device of a storage device A, and in a second storage array, a storage device A' may be used as a backup device of the storage device A, and a storage device B' may be used as a backup device of the storage device B. If the host accesses a storage device (for example, the storage device A), when the storage device A is faulty, the host may restore data from the storage device B, the storage device A', or the storage device B'. Both the first storage array and the second storage array may be implemented by using the storage array 12, and the storage devices A, B, A', and B' may be implemented by using the storage device 13. Referring to FIG. 2, the storage device 13 includes a controller 131 and one or more hard disks 132. The controller 131 includes a communications interface 1311, a processor 1312, a memory 1313, and a bus 1314. The communications interface 1311, the processor 1312, and the memory 1313 are connected to each other by using the bus 1314. The memory 1313 is configured to store data and an operation instruction, and the processor 1312 may perform, by invoking the operation instruction stored in the memory 1313, the SMB service fault processing method shown in FIG. 3 to FIG. 6.
The processor 1312 may be a general purpose processor, including a central processing unit (CPU). The memory 1313 may be a random access memory (RAM).
The SMB service fault processing method provided in the present invention may be implemented based on the foregoing service processing system. A scenario in which the SMB service fault processing method is applied to different storage arrays is first described. Referring to FIG. 3, an embodiment of the SMB service fault processing method provided in the present invention includes the following steps.
Step 301: A first storage device receives a first file open instruction sent by the host 11.
In this embodiment, the first storage array is configured to process a file access request of the host 11. The first storage device may be any storage device that processes an SMB service in the first storage array.
The first file open instruction is used to access a file in the first storage device. When the host 11 sends the first file open instruction to the first storage device, the first storage device generates a file handle according to the first file open instruction, and returns a handle identifier of the file handle to the host 11. The file handle includes handle expectation permission, and the handle expectation permission specifies access permission, exclusive permission, or a range lock for the file. The access permission is permission to read the file, write the file, or delete the file. The exclusive permission indicates that another host cannot perform a read, write, or delete operation on the file. The range lock is used to protect data in a specific area in the file, and the data in the area is readable, writable, or exclusive. After generating the file handle, the first storage device may obtain information such as file permission or a range lock according to the file handle, thereby facilitating handle restoration.
The first file open instruction includes a persistent handle obtaining subcommand, and the first storage device generates a persistent file handle according to the persistent handle obtaining subcommand. Specifically, if the host 11 uses a persistent file handle to access a file, when a storage device is faulty, the host 11 may wait for a period of time during which the host 11 may reconnect to a normal storage device, and send a file open instruction that carries a handle restoration subcommand, so as to restore a service. For example, the persistent handle obtaining subcommand may be a DH2Q subcommand or a DHnQ subcommand, and the handle restoration subcommand is a Dh2C subcommand or a DhnC subcommand.
Step 302: The first storage device generates a file handle according to the first file open instruction, and sends the file handle to a second storage device.
The second storage device is any storage device in the second storage array. As a backup device, the second storage device is configured to process an SMB service when the first storage device is faulty. In a possible manner, a storage device is deployed in the second storage array. A file system of the storage device is a mirror of a file system of the first storage device, and the storage device serves as the second storage device. After receiving file handle backup information, the second storage device saves the file handle backup information. In addition, another storage device may be selected in the second storage array for backup. In a possible implementation, a hash operation is performed on the handle identifier of the file handle to obtain a storage device identifier by means of computation, and a storage device corresponding to the storage device identifier is configured to save the file handle backup information. In another possible manner, a modulo operation is performed on the handle identifier of the file handle, an obtained remainder is a storage device identifier, and a storage device corresponding to the storage device identifier is configured to save the file handle backup information.
When the first storage device executes different types of services, the first storage device may generate a common file handle and a persistent file handle. After generating the file handle, the first storage device checks whether the file handle is a persistent handle; and if the file handle is a persistent handle, the first storage device performs step 303; or if the file handle is not a persistent handle, the first storage device may not send the file handle to the second storage device.
It should be noted that a version number mechanism is used in a handle synchronization process in which the first storage device sends the file handle to the second storage device, and a version number may be configured for both the file and the file handle. A version number is used to distinguish file handles backed up in different periods of time. The second storage device may use a latest version number as a determining basis, and if a version number of a received file handle is not the latest version number, the second storage device may discard the file handle or set the file handle to be invalid.
In addition, the host 11 may send a lock instruction to the first storage device, and the lock instruction is used to set a file lock for a file corresponding to the file handle. The file lock is a range lock, and a file part corresponding to the file lock belongs to readable or writable information. After the file lock is set, file lock backup information is generated, and a correspondence between the file lock backup information and the file handle backup information is established. When the file handle backup information and the file lock backup information are synchronized to the second storage device, the second storage device may obtain the correspondence between the file lock backup information and the file handle backup information.
It can be learned from the above that a lock operation is first performed on the file in the first storage device, and then a step in which the first storage device synchronizes the file lock to the second storage device is performed. If the file lock is faulty to be synchronized, unlocking rollback is performed for the lock operation. In an unlocking process, the file lock backup information is first synchronized, and then unlocking is requested from a distributed lock server in the first storage array. If the file lock backup information is faulty to be synchronized, unlocking is not performed, and the unlocking operation directly is faulty. The distributed lock server may be any storage device in which a distributed lock management application program is deployed, or may be an electronic device that is independently deployed in a storage array and that is configured to manage a distributed lock.
Step 303: The second storage device saves the file handle as a backup handle.
It should be noted that, in a process in which the first storage array synchronizes the file handle or the file lock to the second storage array, if a synchronization error occurs, for example, if the second storage array returns a synchronization error, or the first storage array cannot send the file handle and/or the file lock, the first storage array is set to a state of being unable to switch a service, and then performs resynchronization repair on the backup information. After the repair is completed, synchronization of all backup information is completed, and the first storage array is set to a state of being able to switch a service.
Step 304: The first storage device sends a handle identifier of the file handle to the host 11.
Step 305: When the first storage device is faulty, the second storage device establishes a communication connection to the host 11.
When the first storage array is faulty, after receiving fault information, the second storage array is upgraded from a backup array to a primary service array, and broadcasts an event to each storage device in the second storage array. Any storage device that runs normally is selected from the second storage array, and an IP address of the failed device is activated on the storage device that runs normally. After the IP address is activated, the host 11 establishes a communication connection to the backup device in the second storage array. When obtaining upgrade event information, the second storage device may learn that the first storage device is faulty.
Step 306: The second storage device receives a second file open instruction sent by the host 11. Specifically, after receiving the handle identifier sent by the first storage device, the host 11 generates the second file open instruction that carries the handle identifier. The second file open instruction and the first file open instruction specify the same file. In an optional embodiment, the second file open instruction includes a handle restoration subcommand, and the handle restoration subcommand carries the handle identifier.
Step 307: The second storage device determines, according to the handle identifier in the second file open instruction, backup handle information corresponding to the handle identifier, and opens a file according to the backup handle information.
Step 308: The second storage device sends an operation result to the host 11.
In this embodiment, both the first storage array and the second storage array may be implemented by using the storage array 12, and both the first storage device and the second storage device may be implemented by using the storage device 13. If the host 11 is disconnected from the first storage device, because the second storage device uses the IP address of the first storage device, the host 11 may send a file open instruction to the second storage device, and the second storage device may respond to the file open instruction, obtain the file handle, and then open the file according to the backup handle information, and then return the operation result. When the file is successfully opened, the operation result may be 1, and it indicates that the file is successfully opened. If there is corresponding range lock information in the backup handle information, the distributed lock server restores the range lock according to the range lock information in the backup handle information, and finally returns a restoration result to the host 11.
Based on the embodiment shown in FIG. 3, in an optional embodiment, the foregoing method further includes: after receiving the fault information, sending, by the second storage device, a silence request to the distributed lock server in the second storage array, where the silence request carries file information that needs to be silent; and silently setting, by the distributed lock server, a file corresponding to the backup handle in the second storage device.
In this embodiment, when the first storage array is faulty, after receiving the fault information, the second storage array broadcasts the fault information to each storage device in the second storage array. Storage devices in the second storage array may traverse all backup handles managed by the storage devices, select a backup handle and a backup file lock that are stored for the first storage device, determine, according to the selected backup handle, the file information that needs to be silent, where the file information refers to a file identifier (for example, a file name), and then send the file information to the distributed lock server. The distributed lock server sets a silence identifier for the file corresponding to the backup handle. For example, after the distributed lock server sets the silence identifier for a file text 1 corresponding to a backup handle 1, when receiving another file lock request for the text 1, the distributed lock server does not grant a file lock, so as to avoid a handle restoration fault caused because the file is preempted.
When the service device is faulty, if the service device does not perform a file restoration procedure, a backup handle occupies a storage resource of a backup device, thereby causing a waste of storage resources. To resolve this problem, in another optional embodiment of the present invention, the method further includes: starting, by the second storage device, timing when receiving the fault information sent by the first storage device; and when a timing duration reaches a preset duration, deleting, by the second storage device, the file backup handle, or setting the file backup handle to be invalid.
In this embodiment, a secondary storage device may set an aging duration. When a duration obtained by timing by the secondary storage device reaches the preset aging duration, the secondary storage device may clear a backup handle, so as to save storage resources.
It should be noted that after the fault information is received, processes in which the second storage device sets the silence identifier and the aging duration for the file handle may be performed at the same time.
A scenario in which the SMB service fault processing method is applied to a same storage array is described below. Referring to FIG. 4, another embodiment of the SMB service fault processing method provided in the present invention includes the following steps.
Step 401: A first storage device receives a first file open instruction sent by the host 11.
In this embodiment, the first storage device and a second storage device belong to a same storage array. The first storage device is a service device, and is configured to process an SMB service. The second storage device is any storage device in the storage array. As a backup device, the second storage device is configured to process the SMB service when the first storage device is faulty.
Step 402: The first storage device generates a file handle according to the first file open instruction, and sends the file handle to a second storage device.
Step 403: The second storage device saves the file handle as a backup handle.
For example, referring to FIG. 5, when a host accesses a file file_1 in a storage device A, when receiving a first file open instruction, the storage device A may generate a file handle 1 of the file_1, and synchronize the file handle 1 to a storage device B, and the storage device B saves the file handle 1. The storage device A may further set a file lock (such as a range lock 1 and a range lock 2) for the file_1, and then synchronize the range lock 1 and the range lock 2 to the storage device B.
Step 404: The first storage device sends a handle identifier of the file handle to the host 11.
Step 405: When the first storage device is faulty, the second storage device establishes a communication connection to the host 11.
When the first storage device is faulty, the first storage device may broadcast fault information in the storage array, and the second storage device in the same storage array may detect the fault information by means of listening. The host 11 may activate, on any normal device in the storage array, an IP address of the first storage device, and then establish a communication connection to the second storage device.
Step 406: The second storage device receives a second file open instruction sent by the host 11.
Step 407: The second storage device determines, according to the handle identifier in the second file open instruction, the file handle corresponding to the handle identifier, and opens a file according to backup handle information.
Step 408: The second storage device sends an operation result to the host 11.
For example, referring to FIG. 5, in a process of accessing the file_1 by the host, if the host is disconnected from the storage device A, because the storage device B uses an IP address of the storage device A, the host may send a second file open instruction to the storage device B, and the storage device B may respond to the second file open instruction, obtain the file handle 1, and then open the file according to the file handle 1. The storage device B may obtain the range lock 1 and the range lock 2 according to the file handle 1, and a distributed lock server restores the range locks (that are, the range lock 1 and the range lock 2) of the file, and finally returns a restoration result to the host. It should be noted that after the storage device B is upgraded to a primary service device, a storage device C may be used as a backup device, and after the storage device C is upgraded to the primary service device and the storage device A recovers to normal, the storage device A may be used as a backup device to form a ring backup relationship, so as to ensure reliability of the service processing system.
Based on the embodiment shown in FIG. 4, in an optional embodiment, after step 403, the foregoing method further includes: receiving, by all storage devices, the fault information of the first storage device, and traversing all backup handles generated for the first storage device; and silently setting, by the distributed lock server, files corresponding to the backup handles. In this embodiment, a process in which the distributed lock server silently sets the file corresponding to the backup handle is similar to the process in which the distributed lock server silently sets the file corresponding to the backup handle in the optional embodiment shown in FIG. 3, and details are not described herein.
In an optional embodiment, after step 403, the method further includes: starting, by all the storage devices, timing when receiving the fault information of the first storage device; and when a timing duration reaches a preset duration, deleting, by all the storage devices, the backup handles that are generated for the first storage device and that are managed by all the storage devices, or setting the backup handles to be invalid.
In this embodiment, a secondary storage device may set an aging duration. When a timing duration reaches the preset aging duration, the secondary storage device may clear a backup handle, so as to save storage resources. For ease of understanding, the following uses a specific application scenario to describe in detail the SMB service fault processing method in this embodiment of the present invention.
Referring to FIG. 6, a storage device A belongs to a storage array 1, a storage device A' belongs to a storage array 2, the storage array 2 is a backup array of the storage array 1, the storage device A' is a backup device of the storage device A, a storage device B' is a backup device of the storage device A', a distributed lock server in the storage array 1 is a distributed lock server 1, a distributed lock server in the storage array 2 is a distributed lock server 2, and an IP address of the storage device A is x.x.x.2.
When a host accesses a file file_1 in the storage device A, the host sends a first file open instruction to the storage device A, and a file system of the storage device A accesses the file in response to the first file open instruction. The storage device A generates a file handle 1 of the file_1 according to the first file open instruction, synchronizes the file handle 1 to the storage device A', and returns a handle identifier of the file handle 1 to the host. The storage device A may further generate, according to another file open instruction, a file handle 2 corresponding to a file_2, and synchronize the file handle 2 to the storage device A'.
The host may further initiate a lock instruction for the file_1 to the storage device A, the storage device A initiates a lock request to the distributed lock server 1 according to the lock instruction, and the distributed lock server 1 sets a range lock 1 and a range lock 2 for the file_1. After the locking succeeds, the storage device A synchronizes the range lock 1 and the range lock 2 to the storage device A'.
When the storage array 1 is faulty, after the storage array 2 receives fault information, the storage device A' in the storage array 2 may traverse all backup handles, and select a backup handle (that is, the file handle 1) and backup file locks (that are, the range lock 1 and the range lock 2) that are corresponding to the storage device A, and the distributed lock server 2 sets a silence identifier for the file_1. When receiving the fault information, the storage device A' may further set an aging duration for the file handle 1 of the file_1, and then may set an IP address of the storage device A' to x.x.x.2, and the host establishes a connection to the storage device A'. When the file handle is a persistent handle, the host sends a second file open instruction to x.x.x.2 after waiting for a period of time.
If there is no file lock for the file_1, the storage device A' may obtain the file handle 1 according to the handle identifier that is of the file handle 1 and that is in the second file open instruction, and access the file_1 according to the file handle 1.
If the file_1 is provided with the range lock 1 and the range lock 2, the storage device A' may obtain the file handle 1 and the range lock 1 and the range lock 2 according to the handle identifier that is of the file handle 1 and that is in the second file open instruction, and then the distributed lock server 2 restores the range locks (that are, the range lock 1 and the range lock 2) of the file_1, and returns a restoration result to the host.
A method for processing another file (for example, the file_2) in the storage device A is similar to the method for processing the file_1, and a method for processing a file in another storage device (for example, a storage device B, C, or D) in the storage array 1 is similar to the method for processing the file in the storage device A, and details are not described herein again.
The foregoing describes the SMB service fault processing method in the embodiments of the present invention, and the following specifically describes a service processing system 700 provided in the embodiments of the present invention.
The service processing system includes a host 701 and a storage system 702. The host 701 may implement functions of the host in the foregoing SMB service fault processing method, and the storage system 702 includes at least a first storage device 7021 and a second storage device 7022. The first storage device 7021 may implement functions of the first storage device in the SMB service fault processing method in the embodiments shown in FIG. 3 to FIG. 6, and the second storage device 7022 may implement functions of the second storage device in the SMB service fault processing method in the embodiments shown in FIG. 3 to FIG. 6. The following describes in detail the storage system 702 and the host 701. Referring to FIG. 7, in an embodiment of the storage system 702 provided in the present invention:

the first storage device 7021 is configured to: receive a first file open instruction sent by the host 701, generate a file handle according to the first file open instruction, and send the file handle to the second storage device; and
the second storage device 7022 is configured to save the file handle as a backup handle; and
the second storage device 7022 is further configured to: receive a second file open instruction sent by the host 701; obtain, according to a handle identifier in the second file open instruction, a file handle corresponding to the handle identifier; open a file according to the file handle; and send an operation result to the host 701.

In this embodiment, when the first storage device 7021 is faulty, the second storage device 7022 may take over an SMB service according to the backup file handle, so as to ensure that the SMB service is executed without interruption.
Based on the embodiment shown in FIG. 7, in an optional embodiment, the first file open instruction includes a persistent handle obtaining subcommand. The first storage device 7021 is specifically configured to: generate the file handle according to the persistent handle obtaining subcommand; check whether the file handle is a persistent handle; and if the file handle is a persistent handle, perform the step of sending the file handle to the second storage device 7022.
Based on the embodiment shown in FIG. 7, in an optional embodiment, the second file open instruction includes a handle restoration subcommand, and the handle identifier is carried in the handle restoration subcommand.
In the foregoing optional embodiments, the second storage device 7022 and the first storage device 7021 may belong to a same storage array, or may belong to different storage arrays.
Based on the embodiment shown in FIG. 7, in another optional embodiment of the present invention, referring to FIG. 8, the storage system 702 further includes a distributed lock server 801.
The first storage device 7021 is further configured to: before sending the file handle to the second storage device 7022, receive a lock instruction sent by the host 701; and send a file lock request to the distributed lock server 801 according to the lock instruction. The distributed lock server 801 sets, according to the file lock request, a file lock for the file corresponding to the file handle. The first storage device 7021 sends the file lock to the second storage device 7022; and the second storage device 7022 stores the file lock as a backup file lock.
Based on the embodiment shown in FIG. 7, in an optional embodiment,
the second storage device 7022 is further configured to: after storing the file handle as the backup handle, send a silence request to the distributed lock server 801, where file information carried in the silence request is corresponding to the backup handle; and
the distributed lock server 801 is further configured to silently set, according to the file information, a file corresponding to the backup handle in the second storage device.
Based on the embodiment shown in FIG. 7, in another optional embodiment of the present invention,
the second storage device 7022 is further configured to: determine the backup file lock corresponding to the file handle, and send information about the backup file lock to the distributed lock server 801; and
the distributed lock server 801 is further configured to lock, according to the information about the backup file lock, the file corresponding to the backup handle.
Based on the embodiment shown in FIG. 7, in another optional embodiment of the present invention, the second storage device 7022 is further configured to: start timing when the second storage device 7022 receives fault information sent by the first storage device 7021; and when a timing duration reaches a preset duration, delete the file handle, or set the file handle to be invalid.
Based on the embodiment shown in FIG. 7, in another optional embodiment of the present invention,
the first storage device 7021 is further configured to: before sending the file handle to the second storage device 7022, receive a lock instruction sent by the host 701, and send a file lock request to a first distributed lock server 901 according to the lock instruction; the first distributed lock server 901 sets, according to the file lock request, a file lock for the file corresponding to the file handle; the first storage device 7021 sends the file lock to the second storage device 7022; and the second storage device 7022 stores the file lock as a backup file lock, where the first storage device and the first distributed lock server belong to a same storage array, and the first storage device and the second storage device belong to different storage arrays.
Based on the embodiment shown in FIG. 7, in another optional embodiment,
the second storage device 7022 is further configured to: after storing the file handle as the backup handle, send a silence request to a second distributed lock server, where file information carried in the silence request is corresponding to the backup handle; and
the second distributed lock server 902 is further configured to silently set, according to the file information, a file corresponding to the backup handle in the second storage device, where the second distributed lock server 902 and the second storage device 7022 belong to a same storage array.
Further, in another optional embodiment,
the second storage device 7022 is further configured to: after the second storage device 7022 obtains the file handle corresponding to the handle identifier, determine the backup file lock corresponding to the file handle, and send information about the backup file lock to the second distributed lock server 902; and
the second distributed lock server 902 is further configured to lock the file according to the information about the backup file lock.
The following describes the host 701. Referring to FIG. 10, an embodiment of the host 701 provided in the present invention includes:

a sending module 7011, configured to send a first file open instruction to a first storage device 7021, where the first file open instruction includes a persistent handle request, and
the sending module 7011 is further configured to: when the first storage device 7021 is faulty and an IP address of the first storage device is activated on a second storage device 7022, send a second file open instruction to the second storage device 7022, where the second file open instruction includes a handle identifier, and the handle identifier is used to instruct the second storage device 7022 to obtain a file handle of a file; and
a receiving module 7012, configured to receive an operation result returned by the second storage device 7022.

Based on the embodiment shown in FIG. 10, in an optional embodiment, the first file open instruction includes a persistent handle obtaining subcommand, and the persistent handle obtaining subcommand is used to instruct the first storage device to generate a persistent handle.
Based on the embodiment shown in FIG. 10, in another optional embodiment of the present invention, the second file open instruction includes a handle restoration subcommand, and the handle identifier is carried in the handle restoration subcommand.
Based on the embodiment shown in FIG. 10, in another optional embodiment of the present invention, the sending module 7011 is further configured to send a lock instruction to the first storage device, where the lock instruction is used to set a file lock for a file.
For ease of understanding, the following uses a specific application scenario to describe in detail each part of the service processing system in the embodiments of the present invention.
When the first storage device 7021 belongs to a storage array 1, and the second storage device 7022 belongs to a storage array 2, the storage array 2 is a backup array of the storage array 1, the storage device A' is a backup device of the storage device A, a distributed lock server in the storage array 1 is the distributed lock server 901, a distributed lock server in the storage array 2 is the distributed lock server 902, and an IP address of the storage device A is x.x.x.2.
When a host accesses a file file_1 in the first storage device 7021, the host sends a first file open instruction to the first storage device 7021, and a file system of the first storage device 7021 accesses the file in response to the first file open instruction. The first storage device 7021 generates a file handle 1 of the file_1 according to the first file open instruction, synchronizes the file handle 1 to the second storage device 7022, and returns a handle identifier of the file handle 1 to the host. The first storage device 7021 may further generate, according to another file open instruction, a file handle 2 corresponding to a file_2, and synchronize the file handle 2 to the second storage device 7022.
The host 701 may further initiate a lock instruction for the file_1 to the first storage device 7021, the first storage device 7021 initiates a lock request to the distributed lock server 901 according to the lock instruction, and the distributed lock server 901 sets a range lock 1 and a range lock 2 for the file_1. After the locking succeeds, the first storage device 7021 synchronizes the range lock 1 and the range lock 2 to the second storage device 7022.
When the storage array 1 is faulty, after the storage array 2 receives fault information, the storage device A' in the storage array 2 may traverse all backup handles, and select a backup handle (that is, the file handle 1) and backup file locks (that are, the range lock 1 and the range lock 2) that are corresponding to the storage device A, and the distributed lock server 902 sets a silence identifier for the file_1. When receiving the fault information, the second storage device 7022 may further set an aging duration for the file handle 1 of the file_1, and then may set an IP address of the second storage device 7022 to x.x.x.2, and the host establishes a connection to the second storage device 7022.
When the file handle is a persistent handle, the host sends a second file open instruction to x.x.x.2 after waiting for a period of time.
If there is no file lock for the file_1, the second storage device 7022 may obtain the file handle 1 according to the handle identifier that is of the file handle 1 and that is in the second file open instruction, and access the file_1 according to the file handle 1.
If the file_1 is provided with the range lock 1 and the range lock 2, the second storage device 7022 may obtain the file handle 1 and the range lock 1 and the range lock 2 according to the handle identifier that is of the file handle 1 and that is in the second file open instruction, and then the distributed lock server 902 restores the range locks of the file_1 according to the range lock 1 and the range lock 2, and returns a restoration result to the host. A method for processing another file (for example, the file_2) in the storage device A is similar to the method for processing the file_1, and a method for processing a file in another storage device (for example, a storage device B, C, or D) in the storage array 1 is similar to the method for processing the file in the storage device A, and details are not described herein again.

Claims

A server message block, SMB, service fault processing method, wherein the method is applied to a service processing system (700), and the service processing system (700) comprises a host (11; 701) and at least two storage devices (13; 7021; 7022); and the method comprises:
receiving (301; 401), by a first storage device (7021), a first file open instruction sent by the host (11; 701);

generating (302; 402), by the first storage device (7021), a file handle according to the first file open instruction, sending (302; 402) the file handle to a second storage device (7022), and sending (304; 404) a handle identifier of the file handle to the host (11; 701); and saving (303; 403), by the second storage device (7022), the file handle as a backup handle;

receiving (306; 406), by the second storage device (7022), a second file open instruction sent by the host (11; 701), and determining (307; 407), according to the handle identifier in the second file open instruction, the file handle corresponding to the handle identifier; and

opening (307; 407), by the second storage device (7022), a file according to the file handle, and sending an operation result to the host (11; 701),

wherein before the sending (302; 402) the file handle to a second storage device (7022),

the method further comprises:
receiving, by the first storage device (7021), a lock instruction sent by the host (11; 701);

sending, by the first storage device (7021), a file lock request to a distributed lock server (801) according to the lock instruction;

setting, by the distributed lock server (801) according to the file lock request, a file lock for the file corresponding to the file handle; and

sending, by the first storage device (7021), the file lock to the second storage device (7022); and saving, by the second storage device (7022), the file lock as a backup file lock, wherein the first storage device (7021), the second storage device (7022), and the distributed lock server (801) belong to a same storage array;

or, the method further comprises:
receiving, by the first storage device (7021), a lock instruction sent by the host (11; 701);

sending, by the first storage device (7021), a file lock request to a first distributed lock server (901) according to the lock instruction;

setting, by the first distributed lock server (901) according to the file lock request, a file lock for the file corresponding to the file handle; and

sending, by the first storage device (7021), the file lock to the second storage device (7022); and saving, by the second storage device (7022), the file lock as a backup file lock, wherein

the first storage device (7021) and the first distributed lock server (901) belong to a same storage array, and the first storage device (7021) and the second storage device (7022) belong to different storage arrays.
The method according to claim 1, wherein the first file open instruction comprises a persistent handle obtaining subcommand;
the generating, by the first storage device (7021), a file handle according to the first file open instruction comprises:
generating, by the first storage device (7021), the file handle according to the persistent handle obtaining subcommand; and

the method further comprises:
checking, by the first storage device (7021), whether the file handle is a persistent handle; and if the file handle is a persistent handle, performing a step of sending the file handle to a second storage device (7022).
The method according to claim 1, wherein the second file open instruction comprises a handle restoration subcommand, and the handle identifier is carried in the handle restoration subcommand.
The method according to claim 1, wherein when the first storage device (7021), the second storage device (7022), and the distributed lock server (801) belong to a same storage array, after the saving (303; 403), by the second storage device (7022), the file handle as a backup handle, the method further comprises:
sending, by the second storage device(7022), a silence request to the distributed lock server (801), wherein file information carried in the silence request is corresponding to the backup handle; and

silently setting, by the distributed lock server (801) according to the file information, a file corresponding to the backup handle in the second storage device (7022).
The method according to claim 4, wherein after the determining (307; 407) the file handle corresponding to the handle identifier, the method further comprises:
determining, by the second storage device (7022), the backup file lock corresponding to the file handle, and sending information about the backup file lock to the distributed lock server (801); and

locking, by the distributed lock server (7022), the file according to the information about the backup file lock.
The method according to claim 1, wherein when the first storage device (7021) and the first distributed lock server (901) belong to a same storage array, and the first storage device (7021) and the second storage device (7022) belong to different storage arrays, after the saving (303; 403), by the second storage device (7022), the file handle as a backup handle, the method further comprises: sending, by the second storage device (7022), a silence request to a second distributed lock server (902), wherein file information carried in the silence request is corresponding to the backup handle; and
silently setting, by the second distributed lock server (902) according to the file information, a file corresponding to the backup handle in the second storage device (7022), wherein the second distributed lock server (902) and the second storage device (7022) belong to a same storage array.
The method according to claim 6, wherein after the determining (307; 407) the file handle corresponding to the handle identifier, the method further comprises:
determining, by the second storage device (7022), the backup file lock corresponding to the file handle, and sending information about the backup file lock to the second distributed lock server (902); and

locking, by the second distributed lock server (7022), the file according to the information about the backup file lock.
A storage system, comprising:
a first storage device (7021), configured to: receive a first file open instruction sent by a host (11; 701), generate a file handle according to the first file open instruction, and send the file handle to a second storage device (7022); and

the second storage device (7022), configured to save the file handle as a backup handle, wherein

the second storage device (7022) is further configured to: receive a second file open instruction sent by the host (11; 701), and obtain, according to a handle identifier in the second file open instruction, a file handle corresponding to the handle identifier; and open a file according to the file handle, and send an operation result to the host (11; 701),

wherein the storage system further comprises a distributed lock server (801); and

the first storage device (7021) is further configured to: before sending the file handle to the second storage device (7022), receive a lock instruction sent by the host (11; 701); and send a file lock request to the distributed lock server (801) according to the lock instruction;

the distributed lock server (801) is configured to set, according to the file lock request, a file lock for the file corresponding to the file handle;

the first storage device (7021) is further configured to send the file lock to the second storage device (7022); and

the second storage device (7022)is further configured to store the file lock as a backup file lock, wherein the first storage device (7021), the second storage device (7022), and the distributed lock server (801) belong to a same storage array, or

the storage system further comprises a first distributed lock server (901), the first storage device (7021) and the first distributed lock server (901) belong to a same storage array, and the first storage device (7021) and the second storage device (7022) belong to different storage arrays;

the first storage device (7021) is further configured to: before sending the file handle to the second storage device (7022), receive a lock instruction sent by the host (11; 701); and send a file lock request to the first distributed lock server (901) according to the lock instruction;

the first distributed lock server (901) is configured to set, according to the file lock request, a file lock for the file corresponding to the file handle;

the first storage device (7021) is further configured to send the file lock to the second storage device (7022); and

the second storage device (7022) is further configured to store the file lock as a backup file lock.
The storage system according to claim 8, wherein the first file open instruction comprises a persistent handle obtaining subcommand; and
the first storage device (7021) is specifically configured to generate the file handle according to the persistent handle obtaining subcommand; and
the first storage device (7021) is further configured to: check whether the file handle is a persistent handle; and if the file handle is a persistent handle, perform the step of sending the file handle to the second storage device (7022).
The storage system according to claim 8, wherein the second file open instruction comprises a handle restoration subcommand, and the handle identifier is carried in the handle restoration subcommand.
The storage system according to claim 10, wherein:
when the first storage device (7021), the second storage device (7022), and the distributed lock server (801) belong to a same storage array, the second storage device (7022) is further configured to: after the second storage device (7022) saves the file handle as the backup handle, send a silence request to the distributed lock server (801), wherein file information carried in the silence request is corresponding to the backup handle; and

the distributed lock server (801) is further configured to silently set, according to the file information, a file corresponding to the backup handle in the second storage device (7022).
The storage system according to claim 10, wherein:
when the first storage device (7021), the second storage device (7022), and the distributed lock server (801) belong to a same storage array, the second storage device (7022) is further configured to: determine the backup file lock corresponding to the file handle, and send information about the backup file lock to the distributed lock server (801); and

the distributed lock server (801) is further configured to lock the file according to the information about the backup file lock.